LangGraph

Working experience with LangGraph for constructing stateful, cyclically-structured multi-agent workflows. Enables building agent pipelines with loops, conditional branching, and persistent state checkpointing — going beyond linear LangChain chains to support complex agentic reasoning patterns such as reflection agents, critic-reviewer loops, and tool-use with retry logic. Integrated within agentic AI platforms to coordinate multiple specialized LLM agents with shared memory and streaming execution graphs.